It’s long been said, “Don’t judge a book by its cover.” According to new research by Lisa Feldman Bar­rett, Dis­tin­guished Pro­fessor of Psy­chology at North­eastern Uni­ver­sity, that adage is cer­tainly true when it’s applied to our ability to read emo­tions in a person’s facial expression.

Bar­rett explains that for more than 50 years, sci­en­tists have com­monly thought that there are at least six basic facial expres­sions that indi­cate how a person is feeling. Bar­rett, how­ever, found that many more fac­tors — including body lan­guage, visual scenery, other voices and even a person’s cul­tural ori­en­ta­tion — are essen­tial to per­ceiving emo­tion accurately.

This chal­lenges the long-​​held belief in the sci­ence of psy­chology that faces are the main event,” Bar­rett says.

The find­ings were pub­lished in the October 2011 issue of the review journal Cur­rent Direc­tions in Psy­cho­log­ical Sci­ence. Bar­rett is the lead author, and col­lab­o­rated with North­eastern grad­uate stu­dent Maria Gen­dron, and Batja Mesquita, a researcher at the Uni­ver­sity of Leuven in Belgium.

Bar­rett says, for instance, that a scowl could indi­cate anger, fear or dis­gust. She points to an iconic image of tennis player Serena Williams after she had just defeated her sister, Venus, at the 2008 U.S. Open. By only focusing on Williams’s face, it appears she is screaming in anger. How­ever, zoom out at the entire pic­ture, and she appears ecstatic and is clenching her fist in victory.

The research also found that the con­text of lan­guage aids facial per­cep­tion. She cited a study in which par­tic­i­pants were found to more accu­rately deter­mine emo­tion when asked to choose a term of emo­tion from a pre­de­ter­mined set of words, rather than when par­tic­i­pants were asked to come up with the appro­priate term themselves.

These exam­ples, Bar­rett says, illus­trate that facial expres­sions by them­selves do not broad­cast feel­ings like words on a page. When the con­text is stripped away, a person’s face can broad­cast whether they are in a pos­i­tive or neg­a­tive state or should be approached or avoided — but it doesn’t indi­cate whether that person is angry, sad or afraid.

Bar­rett says these find­ings will have major impli­ca­tions across a broad spec­trum of research areas dealing with emo­tion. For example, secu­rity training methods for law enforce­ment and air­port screeners may have to be reeval­u­ated, given that training is often based on the idea that emo­tions can be read on a person’s face. She also says the find­ings could lead to greater under­standing of the feel­ings of people suf­fering from mental illnesses.

This has real world-​​implications for people to under­stand the limits of their own per­cep­tion,” Bar­rett said. “You are an active archi­tect in the way that you per­ceive the world. You’re not just a sounding board, simply receiving infor­ma­tion from the world and detecting what’s there.”